The link between homelessness or housing instability and HIV is clear and indisputable, particularly among those with long or frequent episodes of homelessness who are also likely to suffer from substance use disorders and serious mental illness. Supportive housing-permanent, subsidized housing with supportive services-has been offered as a structural intervention to reduce HIV infection rates and improve health outcomes of HIV-positive persons. Supportive housing is preferentially offered to the chronically homeless, or those with a homeless period lasting over one year or who have experienced four or more homeless episodes in the last three years and who have one or more qualifying disability including substance use, mental illness and HIV diagnosis. Still, the effects of different supportive housing models, or different program components within these models, and of contextual factors and resident characteristics on health outcomes have not been examined. Nor do we know about the relative cost-effectiveness of the various models. The proposed project will use qualitative and quantitative data to characterize and compare the effectiveness and economic efficiency of at least 20 different supportive housing programs in Chicago, serving over 2000 chronically homeless individuals. We have assembled a strong partnership of researchers at the Center for AIDS Intervention Research (CAIR) and the Center for Housing and Health (CHH), one of the largest providers of supportive housing in Chicago with partnerships with several other supportive housing agencies. Residents in these supportive housing include both HIV positive (approximately half) and HIV negative residents. The project has the following aims: 1) To develop a typology of supportive housing programs using qualitative interviews with program directors and personnel and to explore features of the social and organizational context that may affect the forms supportive housing programs take; 2) To use the developed typology to examine the effectiveness of different types of supportive housing programs on residents' sexual and injection risk, substance use, ART/treatment adherence, and housing satisfaction and stability; 3) To determine per- client housing costs associated with different types of housing programs and whether different types of programs are associated with significantly different per-client external (non-housing) costs; to identify the mot cost-beneficial program types, taking into account both housing and external costs; and to identify program characteristics and client-level factors that predict reduced external costs.